Darwin Lagrangian

1
Lagrangian for a continuous system
Let’s start with an example from mechanics to get the big idea. The physical
system of interest is a string of length  and mass per unit length  fixed at
both ends, and under tension  Choose −axis along the unperturbed string,
and −axis perpendicular to it. When the string is vibrating, its kinetic energy
is:
 =
Z

0
1 2
  =
2
Z

0
1
2
µ


¶2
 =
Z

0
1
(̇)2 
2
To get the potential energy, we use the method of virtual work. The net force
on a string segment has components:
 =  cos 1 −  cos 2 ' 0
and

=  sin 1 −  sin 2 ≈  tan 1 −  tan 2 = 
= 
Ã
2

2
¯
¯ !
 ¯¯
 ¯¯
−
 ¯+  ¯
Then the virtual work is
 =
Z

  =
0
Z


0
2

2
Now integrate by parts, and make use of the fixed end condition:
"
#
Z 
Z
  h 0 2i
0 
0
0
 =   |0 −
 ( )   = −
 ( ) 
2 0
0
R
 then  ·  = − and  = −  ·  Here
Then if  = −∇
 =−
Z
 =

2
Z

2
( 0 ) 
0
The Lagrangian for the string is:
= − =
Z
0

i
1h
2
2
 (̇) −  ( 0 ) 
2
where
i
1h
2
2
 (̇) −  ( 0 )
2
is the Lagrangian density for the string.
L=
1
(1)
The action is
=
Z
 =
Z
2
1
Z

0
i
1h
2
2
 (̇) −  ( 0 ) 
2
Taking the variation of the action, we get
¸
Z 2 Z  ∙
L
L 0 L
 =
 ̇ + 0  +
 
 ̇


1
0
Integrating by parts gives:
Z 2 Z
 =
1

0
∙
¸
L
 L
 L
+
−
−

  ̇
  0

Thus for the action to be an extremum, we need
−
L
 L
 L
+
−
=0
0
  ̇
 

(2)
Using equation (1), we find:


(2̇) +
(−2  0 ) − 0 = 0


or
··
 −   00 = 0
which is the wave equation for the string.
An alternative approach is to write the string displacement as a sum over
normal modes:
X

=
 () sin


Then the Lagrangian density (1) is
L=
XX

̇ ̇ sin




2

sin
−  2   cos
cos





and then the Lagrangian is
=
Z
L
When we integrate over  the only terms that survive are those with  = 
µ
¶
X
2
2
=
 (̇ ) −  2 2 2
(3)
2 

The mode amplitudes  act as the generalized coordinates for the string. Then
Lagrange’s equations are

2
 
··
−
=   −  2 2  = 0
  ̇ 

which is the harmonic oscillator equation with frequency   = ()
2
p
 
2
Lagrangian for the electromagnetic field
Now we want to do a similar treatment for the EM field. We want a Lagrangian
density such that the action
Z
 = L4 
is a Lorentz invariant, and where L is a function of the fields. The "obvious"
invariant to try is
Lguess =   
(Recall this is proportional to  2 −  2  an "energy-like" thing.) Here the components of the potential  are the "normal modes" — they behave like the 
in the previous section. Then Lagrange’s equations (2) are:
L
L

³
´−
=0

  

(4)

To evaluate this, note that
¡
¢
¡
¢
Lguess =    −         −   
and so
Lguess
 (   )
´
³
´
³
¡  
¢ ¡  
¢

 
 
 
 
 

−




−



−




−





+







 
 
 
 
¡  
¢
¡
¢
= (  −   )   −    +    −    (  −   )
= (  −   ) − (  −   ) + (  −   ) − (  −   )
= 4 (  −   ) = 4
=
while
Lguess
=0

So equations (4) become
   = 0
which are Maxwell’s equations in the absence of sources. We can fix up the
Lagrangian by adding the interaction term 1    Thus
L=−
1
1 
  −  
16

With this Lagrangian density
1
L
= − 


and Lagrange’s equations become
−
1 
1
  +  = 0
4

3
or
4


which are the two Maxwell equations that include sources.
   =
3
The Hamiltonian
Now we form the Hamiltonian. First let’s look at the string. Using equation
(3):
X L
̇ − L
 ̇

µ
¶
2
X
2
2
2 (̇ ) −  (̇ ) −  2 2 2
=
2 

2
X
X

=
2 (̇ )2 +  2 2 2 =

2 



=
where  is the total (kinetic plus potential) energy per mode. By analogy, we
get for the EM field system without sources
 
L
   −   L
 (  )
µ
¶
1 
1
= −      −   −
 
4
16
¶
µ
1
1
=
     +     
4
4
=
This tensor is not symmetric, because the first term contains only one half of
the field tensor:    rather than   The conservation laws require that the
energy tensor be symmetric, so we have to modify the result.
4
The energy-momentum tensor
Recall that the field energy density (non-relativistic) is
Poynting theorem may be written
1
8
¡ 2
¢
 +  2 and the
¢
 1 ¡ 2
 ·  
 ×
 +  · 
 =0
 + 2 + ∇
 8
4
4
(5)
We’d like to express this result in covariant form. We obviously need something
quadratic in the fields. For example:
⎞⎛
⎞
⎛
0 − − −


0

⎜
⎜ 
0
 − ⎟
0
−  ⎟
⎟ ⎜ 
⎟
    = ⎜
⎝  −
0
 ⎠ ⎝  
0
− ⎠
  −
0
 − 
0
⎛ 2
 + 2 + 2   −  
  −  
  −  
⎜   −   −2 + 2 + 2 −  −   −  −  
= ⎜
⎝   −   −  −   −2 + 2 + 2 −  −  
  −   −  −   −  −   −2 + 2 + 2
³
´
³
´
³
´
⎞
⎛
 ×

 ×

 ×

2






´
⎟
⎜ ³
⎟
⎜ 
2
2
2


−
+

+

−

−


−

−

×

⎜
 
 
 
  ⎟



⎟
⎜

³
´
= ⎜
⎟
 ×

−  −   −2 + 2 + 2 −  −   ⎟
⎜ 
⎠
⎝ ³
´
 ×

−  −   −  −   −2 + 2 + 2

⎞
⎟
⎟
⎠

Now we’d like the (0,0) component to be the
¡ energy¢ density. We can get that
if we add the tensor 14      = 12    2 −  2  Then
¾
½
1
1
    +     
Θ =
4
4
³
´
³
´
³
´
⎛
2
2
 ×

 ×

 ×

 2 +  −



2



´
⎜ ³
⎜ 
2
2
2  2 − 2


−  +  +  − 2
−   −  
−   −  
×
1 ⎜
⎜ ³
´
=
2
2
⎜
2
2
2

−
 ×

4 ⎜ 
−   −  
−  +  +  − 2
−   −  
⎝ ³
´
2
2
 ×

−   −  
−   −  
− 2 + 2 + 2 −  −

2

³
´
³
´
³
´
⎛  2 + 2
⎞
 ×

 ×

 ×




2



´
⎜ ³
⎟
⎜ 
⎟
 2 + 2
2
2
 ×

−

−

−

−


−

−


⎟










1 ⎜
2
⎜
⎟

³
´
=
2
2
⎜
⎟
+

2
2
 ×

4 ⎜ 
−  −  
−

−

−

−


⎟
 
 


2
⎝ ³
⎠
´
2
2
 +
2
2
 ×

−  −  
−  −  
−

−




2

Then
0
 Θ
 2 + 2 
=
+∇

8
Ã
 ×


4
!
 = 1   0 
which is part of equation (5). On the right hand side we need 1  · 

Thus we have the relation
 Θ0 =
5
1
  0

⎞
⎟
⎟
⎟
⎟
⎟
⎟
⎠
and so we guess that the full set of conservation laws are given by:
 Θ =
1
  

I leave it to you to show that the  =  components give momentum conservation
(Jackson equation 6.122).
5
Angular momentum
Cross products are not proper vectors. They are pseudo-vectors because they
do not transform properly under reflections. Thus it is usually better to express
 =  × ) as antisymmetric
quantities such as angular momentum of a particle (
tensors. For example the tensor
X
 =
(  −   )
particles

has three independent compomemts: the components of the vector 
Extending this idea, let’s look at the tensor
X ¡
¢
  =
  −  
particles
The 3×3 spacelike part is the tensor  and thus represents the angular
momentum of the system. In addition:
´
X³ 
 0 =
 − 

where  is the energy of the particle. Conservation of angular momentum for
the system is expressed as  = constant. Thus we conjecture that the full
conservation law is   = constant. (Or equivalently    = 0) This gives
for the ( 0) component:
´
X³ 
 0 =
 −  = constant

P
Now if we divide through by
 we get
P 
P 


P = 2  P


The term on the left hand side is the position of the center of mass,
P

 = P

while the term on the right hand side is the CM velocity times 
P

 = P

6
an eminently sensible result.
To get the equivalent result for the EM field we form the tensor
  = Θ  − Θ 
and then the conservation laws should be given by:
   = 0
Taking  = 0 gives the CM motion as above.
6
The Darwin Lagrangian
The analysis above is for source-free fields. We might attempt to add the freeparticle Lagrangian to get a complete description of the particle-plus-field system, but this approach fails because of retardation effects. (The fields propagate
at the speed of light.) We can calculate a complete Lagrangian in a single reference frame, inlcuding relativistic effects up to order  2 = ()2 
Let’s start with a 2-particle system. Both particles produce fields and both
can move under the influence of those fields. The interaction term for charge 1
interacting with the fields due to 2 is
µ
¶
³
´
1
1
1 

=
)





−
(6)
=

1 
(


·

1
2
1
2
2
2
2



If we now work in a single reference frame and use the coordinate time rather
than proper time as our time variable, we should drop the factor  We want to
evaluate this expression to second order in  If we work in Coulomb gauge,
 to first order since it appears
the potential 2 = 2  is exact. We only need 
in combination with 1  This means we can ignore retardation effects. Then:
2 ' 1


Z

 0
| − 0 |
where the transverse current is

Therefore
Z 0
1 
∇ · 22  (0 − 2 ) 0


=  −  = 22  ( − 2 ) −
∇

4
| − 0 |
2  2 · ( − 2 )
= 22  ( − 2 ) −
∇
4
| − 2 |3
 = 2 2 − 2

 
4
Z
0
1
 0 2 · ( − 2 )  0
∇
3
| − 0 |
|0 − 2 |
7
Let’s look at the integral. First make a change of origin. Let  = 0 − 2 and
with  =  − 2  we get
Z
Z
0
1
1  2 ·  3
 0 2 · ( − 2 )  0 =
∇
∇ 3  
3
0
0
| −  |
| − |

| − 2 |
¯
µ
¶
Z
¯
1
2 ·  3
1 2 ·  ¯

−
∇
 
=
| − | 3 ¯ at ∞
 − 
3
Z
1 2 · ̂ 3

= ∇
 
| − | 2
Z X
∞



 () 2 · ̂ 
= ∇
+1
=0 
where we have put the polar axis for  along  Now
2 · ̂ = 2 · (̂ cos  + sin  (̂ cos  + ̂ sin ))
Integration over  renders the − and −components zero.
Next we make use of the orthogonality of the  ()  noting that cos  =
1 ()  Only  = 1 survives the integration over  We obtain:
Z ∞
 2

integral = 2∇
2 · ̂ 
2 3

0
µZ 
¶
Z ∞


4 
=
 +

∇ 2 · ̂
3
2
2

µ 0 ¶
µ
¶
4 
1


=
∇ 2 · ̂
+ 1 = 2∇ 2 ·
3
2

And thus
µ
¶
2 2

2

=
−
2∇ 2 ·
 
4

¶¸
∙
µ
2 2
1 2 2 · 
=
−
− 2 ̂
 
2 

2
=
[2 + (2 · ̂) ̂]
2
Then the interaction term for 2 particles (equation 6 with the  dropped) is:
Ã
!
¶
µ
2
1 · 
2 1 2
1 2 −
= 1
−
·
[2 + (2 · ̂) ̂]


 2
¾
½
1
2
= 1
1 − 2 [1 · 2 + (2 · ̂) (1 · ̂)]

2
2

Adding this term to the kinetic energy (Lagrangian notes pg 5), we have the
Darwin Lagrangian for a collection of charged particles:
r
¾
½
1X
 2 X  
1
2
 = −
  1 − 2 −
1 − 2 [ ·  + ( · ̂) ( · ̂)]
2 


2

8
To be consistent, we should evaluate the first term to second order in i.e.
¢12
¡ ¢ 4
¡
2
' 1 − 12 2 + 12 − 12 2
1 − 2 2
4  Finally, dropping the constant leading
term which is irrelevant, we have
¾
X   ½
1X
1 X
1
2
4
  + 2
  −
1 − 2 [ ·  + ( · ̂) ( · ̂)]
 =
2 
8 

2

correct to second order in 
9